Intraductal papillary mucinous neoplasms (IPMNs) are classified into three types according to tumor localization: main duct IPMN (MD-IPMN), branch duct IPMN (BD-IPMN), and mixed-type IPMN (MT-IPMN), but their natural history is unclear.¹ Pancreatic MD- and MT-IPMNs have a higher malignancy rates than BD-IPMNs.² Mural nodules are predictors of malignancy in IPMNs of the pancreas.^2,3 Most guidelines recommend surgery of all MD- and MT-IPMNs in suitable patients regardless of the presence of mural nodules. Even though surgical resection is the best way to treat and cure IPMNs, it can also lead to significant morbidity and mortality.^4-6 However, IPMNs occur predominantly in elderly patients, in which operative risk can be high.⁷ Despite a consensus recommendation for surgery, it is unclear whether all patients with MD- and MT-IPMN will profit from surgery.

Most studies that show a strong association between enhancing mural nodules (EMNs) and malignancy have been conducted in patients with BD-IPMN.^8-10 The accuracy of histological diagnosis of mural nodules varies. However, there is little evidence on the malignancy risk factors of MD- and MT-IPMNs with EMNs in just the main pancreatic duct (MPD). Therefore, the aim of this study was to establish the clinical and morphological characteristics associated with malignancy in patients with MD- and MT-IPMNs with EMNs only in the MPD by contrast-enhanced magnetic resonance imaging (CE-MRI).

1. Patients

We conducted a retrospective review of patients with pathologically confirmed MD- and MT-IPMNs after pancreatic resection at two academic medical centers between January 2002 and December 2019. The study protocol was approved by each institution (IRB number: WKUH 2022-07-013) and the informed consent was waived. During the study period, we identified 344 patients who underwent surgery of MD- or MT-IPMNs. Of these, 294 patients were excluded for no EMN in the MPD (n=226), definitive mural nodules in the BD (n=46), synchronous pancreatic solid mass (n=7), and preoperative MRI not performed (n=15). Finally, 50 patients were enrolled in the study (Fig. 1). We evaluated their demographic data, preoperative symptoms, acute pancreatitis history, preoperative serum carbohydrate antigen 19-9 (CA19-9) level, carcinoembryonic antigen level, radiological features, and pathological surgical results.

Figure 1. Flowchart of the study population.
MD, main duct; IPMN, intraductal papillary mucinous neoplasm; MRI, magnetic resonance imaging; MPD, main pancreatic duct.

According to the 2017 international consensus guidelines (ICG), MD-IPMNs were defined as dilated MPD >5 mm on preoperative radiologic imaging, including CE computed tomography and MRI, without other natures of obstruction. MT-IPMNs were defined as dilated MPD >5 mm communicating with a dilation of the BD (>5 mm diameter).¹ Elevated serum CA19-9 and carcinoembryonic antigen levels were defined as >37 U/mL and >5.0 ng/mL, respectively.

2. Preoperative radiologic studies

All MRI examinations were performed using the 1.5 Tesla or 3.0 Tesla Magnetom Avanto system. Dynamic CE-MRIs were obtained using T-1 weighted volumetric sequences acquired before and after injection of 0.1 mmol/kg gadoxetic acid (Primovist; Bayer Healthcare, Berlin, Germany) or gadopentetic acid (Magnevist; Bayer Healthcare).

Two gastroenterologists (T.H.K and T.J.S, with 23 years and 14 years of pancreatobiliary experience, respectively) independently reviewed MRI images including EMN size. They were blinded to other clinical information and pathologic results. In cases of discrepancy between the two reviewers, a third reviewer (H.K.C., with 9 years of experience in the pancreatobiliary field) was consulted. Afterward, the three reviewers agreed on a final decision.

For preoperative radiologic evaluation, we reviewed the MRIs performed on all subjects 3 months before surgery. The presence and size of EMNs, cyst size, MPD diameter, and morphological pattern of MPD dilatation were retrospectively evaluated. On MRI, EMNs were defined as solid nodules that showed enhancement only in the MPD (Fig. 2). EMN size, largest cyst diameter, and maximal MPD diameter were measured in every plane of the MRI. The MPD morphological pattern was classified into two types according to the scale of dilatation: a diffuse type with MPD dilatation two-thirds or more of the entire MPD length and a segmental type with MPD dilation less than two-thirds of the total MPD length. According to the 2017 ICG, dilated MPD was classified based on the maximum diameter (5–9 mm vs ≥10 mm).

Figure 2. Main duct intraductal papillary mucinous neoplasm in a 67-year-old man. (A) Contrast-enhanced magnetic resonance image shows an enhanced mural nodule (10 mm) (arrow) in the dilatated main pancreatic duct of the pancreatic head. (B) Magnetic resonance cholangiopancreatography shows a round signal avoiding a lesion (arrow) in the dilatated main pancreatic duct of the pancreatic head. (C) Gross findings of the surgical specimen show a mural nodule in the main pancreatic duct of the pancreas in the presence of invasive carcinoma.

3. Pathological analysis

Pathological diagnosis of the surgical specimen was made by each pathologist at each institution based on the classification of the World Health Organization at the time of diagnosis. MPD involvement in all resected specimens was pathologically established, and histology showed a neoplasm with columnar, mucin-producing epithelium, with or without papillary growth, within the MPD epithelial lining. In this study, resected specimens were classified into three groups: invasive carcinoma derived from IPMNs, high-grade dysplasia (HGD), and low-grade dysplasia. HGD and invasive carcinoma were designated as malignant and low-grade dysplasia as benign.

4. Statistical analyses

Statistical analyses were conducted using IBM SPSS Statistics for Windows, version 22 (IBM Corp., Armonk, NY, USA). Continuous variables are shown as median with interquartile range, and the Wilcoxon rank-sum test was executed after the normality test (Kolmogorov-Smirnov test). Categorical variables were compared using the chi-square and Fisher exact probability tests. Statistical significance was considered at p<0.05. Receiver operating characteristic curve analysis was used to determine the optimal cutoff values for EMN size and MPD diameter in malignant IPMNs. Multivariate analysis using logistic regression models was used to estimate the effects of possible predictive factors on malignant MD- and MT-IPMNs. When conducting multivariate analysis, it would be common to change in several predictive variables’ significance level. According to the several studies, variables have been included in the multivariate analysis when a p-value <0.2 in observed in univariate analysis.

1. Baseline characteristics

Fifty patients who met the inclusion criteria were enrolled at each participating center. Their baseline characteristics are shown in Table 1. The median age was 66.5 years (interquartile range, 59 to 71 years), and 32 patients (64%) were men. Most patients were asymptomatic, but six (12%) presented with abdominal pain, and four (8%) experienced weight loss. Acute pancreatitis occurred in two patients (4%). Of the 50 patients, 29 (58%) had MD-IPMNs, and 21 (42%) had MT-IPMNs. The malignancy rate was 62% (17 HGD and 14 invasive carcinoma cases). There were no significant differences in age, sex, proportion of MD- or MT-IPMNs, cyst size, or median MPD diameter between the benign and malignant groups. Clinical symptoms were more common in the malignant group, including weight loss or abdominal pain, diffuse MPD dilatation, and abnormal serum CA19-9 level. The size of the EMN in the MPD was statistically larger in the malignancy group than in the benign group (11.00±11.08 mm vs 7.13±3.62 mm, p=0.019).

Table 1 . Characteristics of Patients with Main Duct and Mixed-Type IPMNs of the Pancreas with EMNs Only in the MPD on Magnetic Resonance Imaging.

Characteristic	Total (n=50)	Benign (n=19)	Malignancy (n=31)	p-value
Age, yr	66.5 (59.0–71.0)	64 (58.0–70.0)	67 (60.0–72.0)	0.394*
Male sex	32 (64.0)	13 (68.4)	19 (61.3)	0.610
Episode of acute pancreatitis	2 (4.0)	1 (5.3)	1 (3.2)	>0.999†
Presence of symptoms
Asymptomatic	10 (20.0)	3 (15.8)	7 (22.6)	0.722†
Body weight loss	4 (8.0)	1 (5.3)	3 (9.7)	<0.001†
Abdominal pain	6 (12.0)	2 (10.5)	4 (12.9)	<0.001†
Pathological diagnosis				<0.001
Low grade	19 (38.0)	19 (100.0)	0
High grade	17 (34.0)	0	17 (54.8)
Invasive cancer	14 (28.0)	0	14 (45.2)
Background data on imaging findings
Main duct/mixed type	29/21	9/10	20/11	0.233
MPD diameter, mm	11 (7–15)	8 (7–13)	12 (7–16)	0.298*
Cyst size, mm	0 (0–5)	1 (0–4)	0 (0–5)	0.436*
Morphological classification of MPD
Dilatation extent (diffuse/segmental type)	43/7	14/5	29/2	0.054†
Dilatation degree (≥10 mm/5–9 mm)	28/22	9/10	19/12	0.336
MPD diameter, mm	11 (7.0–15.0)	8 (7.0–12.5)	12 (7.2–15.5)	0.128
Size of EMNs in MPD, mm	11.00±11.08	7.13±3.62	13.37±1.32	0.019
Serum tumor marker
Serum level of CEA, ng/mL	1.6 (1.0–3.0)	2 (1.0–3.0)	1.4 (1.0–3.0)	0.443*
Serum level of CA19-9, U/mL	16.5 (10.0–34.0)	11.7 (9.0–25.0)	23.0 (10.0–56.0)	0.048*

Data are presented as median (interquartile range), number (%), or mean±SD..

IPMN, intraductal papillary mucinous neoplasm; EMNs, enhancing mural nodules; MPD, main pancreatic duct; CEA, carcinoembryonic antigen; CA19-9, carbohydrate antigen 19-9..

*Wilcoxon rank-sum test; ^†Fisher exact test..

Pancreaticoduodenectomy was performed in 22 patients (44%), distal pancreatectomy in 20 patients (40%), and total pancreatectomy in eight patients (1.6%). The final pathological diagnoses of the 50 patients included low-grade dysplasia in 19 patients (38.0%), HGD in 17 (34%), and invasive IPMNs in 14 (28%) (Fig. 3). The malignancy rate in all 50 patients with MD- or MT-IPMNs was 62% (31/50).

Figure 3. Cutoff values for enhancing mural nodule size and main pancreatic duct diameter. (A) Cutoff point for mural nodule size: 4.99 mm (sensitivity, 93.55%; specificity, 52.63%; Youden index, 46.18%). (B) Cutoff point for MPD diameter: 8.50 mm (sensitivity, 64.52%; specificity, 52.63%; Youden index, 17.15%).
ROC, receiver operating characteristic; AUC, area under the curve; MPD, main pancreatic duct.

2. Determination of cutoff values of EMN size and MPD diameter

The receiver operating characteristic curve was used to identify the cutoff values of EMN size and MPD diameter for predicting malignancy. The results showed values for EMN size and MPD diameter of 5 mm (sensitivity, 93.55%; specificity, 52.63%; Youden index, 46.18%; area under the curve, 0.753) and 8.5 mm (sensitivity, 64.52%; specificity, 52.63%; Youden index, 17.15%; area under the curve, 0.590), respectively (Fig. 3).

3. Malignancy predictors for all patients with MD- or MT-IPMNs

Although age, sex, acute pancreatitis, clinical symptom incidence, and abnormal serum carcinoembryonic antigen level did not differ significantly between malignant and benign lesions, serum CA19-9 level was significantly different between the two groups. In univariate analysis, increased serum CA19-9 level, diffuse MPD dilation, MPD >8.5 mm, and EMN >5 mm were associated with malignancy (Table 2). In multivariate analysis, only EMN >5 mm (odds ratio, 27.69; 95% confidence interval, 2.75 to 278.73; p=0.005) was statistically significant (Table 2).

Table 2 . Univariate and Multivariate Analyses for Risk Factors for Malignant MD- or MT-IPMNs.

Factor		Univariate analysis			Multivariate analysis*
		OR (95% CI)	p-value		OR (95% CI)	p-value†
Age		1.01 (0.96–1.06)	0.660
Sex	Male	0.73 (0.22–2.45)	0.611
	Female	1 (reference)
Episode of acute pancreatitis	Yes	0.60 (0.04–10.20)	0.724
	No	1 (reference)
Presence of symptoms: asymptomatic	Yes	1.56 (0.35–6.92)	0.562
	No	1 (reference)
Presence of symptoms: body weight loss	Yes	1.93 (0.19–20.01)	0.582
	No	1 (reference)
Presence of symptoms: abdominal pain	Yes	1.26 (0.21–7.64)	0.802
	No	1 (reference)
Serum level of CA19-9		1.03 (0.99–1.07)	0.118		1.05 (0.99–1.11)	0.128
Serum level of CEA		1.02 (0.81–1.28)	0.891
MD-IPMN (vs MT-IPMN)	MD-IPMN	2.02 (0.63–6.46)	0.236
	MT-IPMN	1 (reference)
Pattern of MPD dilatation	Segmental	0.19 (0.03–1.12)	0.067		0.65 (0.05–8.01)	0.737
	Diffuse	1 (reference)			1 (reference)
MPD diameter	>8.5 mm	2.33 (0.72–7.55)	0.157		0.34 (0.05–2.31)	0.271
	≤8.5 mm	1 (reference)			1 (reference)
Size of EMNs in MPD	>5 mm	16.11 (2.97–87.52)	0.001		27.69 (2.75–278.73)	0.005
	≤5 mm	1 (reference)			1 (reference)

MD, main duct; MT, mixed type; IPMN, intraductal papillary mucinous neoplasm; OR, odds ratio; CI, confidence interval; CA19-9, carbohydrate antigen 19-9; CEA, carcinoembryonic antigen; MPD, main pancreatic duct; EMNs, enhancing mural nodules..

*Only variables significant at p≤0.20 in univariate analysis were included; ^†p-value of Hosmer and Lemeshow goodness-of-fit test is 0.662..

Most guidelines recommend surgery of all MD- and MT-IPMNs in suitable patients. However, there is little evidence regarding the malignancy risk of MD- and MT-IPMNs with EMNs in the MPD. Additionally, these types of IPMNs may contain a heterogeneous group of intraductal mucin-producing neoplasms, representing a typical adenoma-to-carcinoma sequence. Therefore, clinicians should focus on identifying potentially reliable predictors of IPMN with invasive cancer and HGD. This study found that an EMN size >5 mm may be related to malignancy in patients with MD- and MT-IPMNs with EMNs in the MPD on CE-MRI.

Mural nodules in the cyst and MPD have been identified as independent criteria for predicting malignant lesions when treating IPMNs.^2,11 A recent meta-analysis emphasized the key role of vascularized solid nodules.¹¹ In this study, we included resected MD- and MT-IPMNs with mural nodules. A mural nodule size >5 mm in the MPD was an independent risk factor for malignancy in the MD- and MT-IPMNs of the pancreas. In the literature, the size cutoff for solid nodules that can predict IPMN malignancy has been debated. Uehara et al.¹² demonstrated that mural nodules >10 mm were independent factors significantly associated with malignancy in BD-IPMNs. Kawada et al.¹³ found that 70% of patients with a >10 mm wall nodule had HGD.¹³ However, an EMN >5 mm in the BD- and MT-IPMNs was an independent predictive factor for HGD. The 2017 ICG and 2018 European guidelines emphasize the importance of size >5 mm in the EMNs of BD-IPMN as a high-risk signal and an absolute indication for surgical resection. However, there is lack of evidence regarding the malignancy risk of MD- and MT-IPMNs with mural nodules in the MPD. The results of this study support these guidelines. Rather than immediate resection of all MD-IPMNs with a mural nodule in the MPD, they might be managed by surgery or observation according to the presence of EMN >5 mm.

IPMNs with dilated MPD in the pancreas have more aggressive nature and higher malignant rates than do IPMNs without MPD dilation. However, MPD dilation may often result from the mucous made by the underlying tumor process in the duct walls or a neoplasm diffusely involving the MPD. MPD dilation represents a wide range of conditions, from IPMNs to other benign (ductal hypertension caused by mucin hypersecretion or mucin plugs and chronic obstructive pancreatitis) as well as malignant entities (solid tumors with retrograde MPD dilation). The probability of malignancy increases when the MPD is >10 mm.² However, not all MD-IPMNs with MPD >10 mm is malignant. Hackert et al.¹⁴ showed that the malignancy rates of IPMNs with MPD diameters of 5–9 mm and ≥10 mm were 59% and 73%, respectively. Sugimoto et al.¹⁵ reported similar findings, with malignant rates of 53.7% and 67.7%, respectively. Several papers based on surgical series reported cutoff values of 8, 10, 12, and 15 mm with sensitivities of 86%, 50%, 74%, and 86% and specificities of 69%, 88%, 90%, and 78%, respectively. When we determined the cutoff level of 8.5 mm for MPD diameter by receiver operating characteristic analysis, it was not a significant factor associated with malignancy in MD- and MT-IPMNs with mural nodules in the MPD. The size of the MPD may be related to the malignant potential of IPMNs, but is no consistent cutoff value that predicts malignancy.

In this study, diffuse-type MPD dilatation on preoperative radiological findings was not an independent risk factor for malignancy, although this morphology was significantly higher than in the malignant group than in the benign group by univariate analysis. Roch et al.¹⁶ reported that diffuse-type MPD dilatation was independently related to the malignancy in MT-IPMNs. Diffuse dilated MPD may indicate broader disease involvement.

The limitations of this study are its retrospective design and small cases. Prospective studies with more data are needed to confirm these findings. Additionally, endoscopic ultrasound may be the most accurate method for measuring mural nodules size in pancreatic cystic neoplasms. However, we only utilized mural nodule size as measured by MRI, which could have introduced a diagnostic bias. Furthermore, only patients with surgery of MD- and MT-IPMNs were enrolled in this study; therefore, it did not cover the natural course of IPMN. Finally, the long study period was both a strength and a limitation of our study. Compared with other studies of MD- and MT-IPMNs, the number of enrolled patients was relatively large, but there may have been missing data. Therefore, national observational cohort studies are needed to thoroughly address many investigative questions in MD- and MT-IPMNs.

In conclusion, EMN size >5 mm may be related to malignancy in patients with MD- and MT-IPMNs with mural nodules in the MPD on CE-MRI. These findings support the MD- and MT-IPMNs management algorithms in the 2017 ICG. The mural nodule size in MPD may be helpful for planning personalized surveillance of BD-IPMNs. These results should be validated in a large-scale, prospective study.

This study was supported by Wonkwang University 2022.

No potential conflict of interest relevant to this article was reported.

Study concept and design: T.H.K., M.H.K. Data acquisition, analysis and interpretation: H.K.C., K.H.Y., J.S.H. Drafting of manuscript: T.H.K., H.K.C. Critical revision of the manuscript for important intellectual content: T.J.S., M.H.K., E.K.C. Statistical analysis: T.H.K. Obtained funding: T.H.K. Supervision: T.H.K, M.H.K. Approval of final manuscript: all authors.